Fig. 2From: A robust nonlinear low-dimensional manifold for single cell RNA-seq dataComparison of manifold learning methods on 11 neural and blood cell populations [33]. a Average adjusted rand score (ARS; left) and normalized mutual information (NMI; right) of ten k-means cluster labels versus available cell type labels with respect to the number of latent dimensions. b Three dimensional latent mappings from tGPLVM, PCA, and ZIFA colored by inferred cell type label. t-SNE (not pictured) collapses in three dimensionsBack to article page